Currently there is a broken abstraction layer which has been used as an overlay for logging and telemetry data. This does not work as expected and is not an effective solution. Instead, I believe we should simply use open telemetry directly as the abstraction layer given its extensibility and compatibility with other logging and telemetry solutions.
The best way to do this without breaking existing code is to find all occurrences of the abstraction layer and incrementally replace them with the direct open telemetry implementation. This should be done service by service in the following order:
- Core
- Integration manager
- Subject
- Service manager
- Default API
- Default facade
Once this is done, a default open telemetry configuration with reasonable defaults and centralized logging configuration.
Currently there is a broken abstraction layer which has been used as an overlay for logging and telemetry data. This does not work as expected and is not an effective solution. Instead, I believe we should simply use open telemetry directly as the abstraction layer given its extensibility and compatibility with other logging and telemetry solutions.
The best way to do this without breaking existing code is to find all occurrences of the abstraction layer and incrementally replace them with the direct open telemetry implementation. This should be done service by service in the following order:
Once this is done, a default open telemetry configuration with reasonable defaults and centralized logging configuration.